[PATCH v3 0/5] block: remove gfp_mask for blkdev_zone_mgmt()

[PATCH v3 0/5] block: remove gfp_mask for blkdev_zone_mgmt()

[Johannes Thumshirn]

Fueled by the LSFMM discussion on removing GFP_NOFS initiated by Willy,
I've looked into the sole GFP_NOFS allocation in zonefs. As it turned out,
it is only done for zone management commands and can be removed.

After digging into more callers of blkdev_zone_mgmt() I came to the
conclusion that the gfp_mask parameter can be removed alltogether.

So this series switches all callers of blkdev_zone_mgmt() to either use
GFP_KERNEL where possible or grab a memalloc_no{fs,io} context.

The final patch in this series is getting rid of the gfp_mask parameter.

Link: https://lore.kernel.org/all/ZZcgXI46AinlcBDP@casper.infradead.org/

---
Changes in v3:
- Fix build error after rebase in dm-zoned-metadata.c
- Link to v2: https://lore.kernel.org/r/20240125-zonefs_nofs-v2-0-2d975c8c1690@wdc.com

Changes in v2:
- guard blkdev_zone_mgmt in dm-zoned-metadata.c with memalloc_noio context
- Link to v1: https://lore.kernel.org/r/20240123-zonefs_nofs-v1-0-cc0b0308ef25@wdc.com

---
Johannes Thumshirn (5):
      zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call
      dm: dm-zoned: guard blkdev_zone_mgmt with noio scope
      btrfs: zoned: call blkdev_zone_mgmt in nofs scope
      f2fs: guard blkdev_zone_mgmt with nofs scope
      block: remove gfp_flags from blkdev_zone_mgmt

 block/blk-zoned.c              | 19 ++++++++-----------
 drivers/md/dm-zoned-metadata.c |  5 ++++-
 drivers/nvme/target/zns.c      |  5 ++---
 fs/btrfs/zoned.c               | 35 +++++++++++++++++++++++++----------
 fs/f2fs/segment.c              | 15 ++++++++++++---
 fs/zonefs/super.c              |  2 +-
 include/linux/blkdev.h         |  2 +-
 7 files changed, 53 insertions(+), 30 deletions(-)
---
base-commit: 615d300648869c774bd1fe54b4627bb0c20faed4
change-id: 20240110-zonefs_nofs-dd1e22b2e046

Best regards,
-- 
Johannes Thumshirn <johannes.thumshirn@wdc.com>

[PATCH v3 1/5] zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call

[Johannes Thumshirn]

Pass GFP_KERNEL instead of GFP_NOFS to the blkdev_zone_mgmt() call in
zonefs_zone_mgmt().

As as zonefs_zone_mgmt() and zonefs_inode_zone_mgmt() are never called
from a place that can recurse back into the filesystem on memory reclaim,
it is save to call blkdev_zone_mgmt() with GFP_KERNEL.

Link: https://lore.kernel.org/all/ZZcgXI46AinlcBDP@casper.infradead.org/
Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
---
 fs/zonefs/super.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)

diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index 93971742613a..63fbac018c04 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -113,7 +113,7 @@ static int zonefs_zone_mgmt(struct super_block *sb,
 
 	trace_zonefs_zone_mgmt(sb, z, op);
 	ret = blkdev_zone_mgmt(sb->s_bdev, op, z->z_sector,
-			       z->z_size >> SECTOR_SHIFT, GFP_NOFS);
+			       z->z_size >> SECTOR_SHIFT, GFP_KERNEL);
 	if (ret) {
 		zonefs_err(sb,
 			   "Zone management operation %s at %llu failed %d\n",

-- 
2.43.0

[PATCH v3 2/5] dm: dm-zoned: guard blkdev_zone_mgmt with noio scope

[Johannes Thumshirn]

Guard the calls to blkdev_zone_mgmt() with a memalloc_noio scope.
This helps us getting rid of the GFP_NOIO argument to blkdev_zone_mgmt();

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
---
 drivers/md/dm-zoned-metadata.c | 5 ++++-
 1 file changed, 4 insertions(+), 1 deletion(-)

diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
index fdfe30f7b697..165996cc966c 100644
--- a/drivers/md/dm-zoned-metadata.c
+++ b/drivers/md/dm-zoned-metadata.c
@@ -1655,10 +1655,13 @@ static int dmz_reset_zone(struct dmz_metadata *zmd, struct dm_zone *zone)
 
 	if (!dmz_is_empty(zone) || dmz_seq_write_err(zone)) {
 		struct dmz_dev *dev = zone->dev;
+		unsigned int noio_flag;
 
+		noio_flag = memalloc_noio_save();
 		ret = blkdev_zone_mgmt(dev->bdev, REQ_OP_ZONE_RESET,
 				       dmz_start_sect(zmd, zone),
-				       zmd->zone_nr_sectors, GFP_NOIO);
+				       zmd->zone_nr_sectors, GFP_KERNEL);
+		memalloc_noio_restore(noio_flag);
 		if (ret) {
 			dmz_dev_err(dev, "Reset zone %u failed %d",
 				    zone->id, ret);

-- 
2.43.0

[PATCH v3 3/5] btrfs: zoned: call blkdev_zone_mgmt in nofs scope

[Johannes Thumshirn]

Add a memalloc_nofs scope around all calls to blkdev_zone_mgmt(). This
allows us to further get rid of the GFP_NOFS argument for
blkdev_zone_mgmt().

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
---
 fs/btrfs/zoned.c | 33 ++++++++++++++++++++++++++-------
 1 file changed, 26 insertions(+), 7 deletions(-)

diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index 168af9d000d1..05640d61e435 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -824,11 +824,15 @@ static int sb_log_location(struct block_device *bdev, struct blk_zone *zones,
 			reset = &zones[1];
 
 		if (reset && reset->cond != BLK_ZONE_COND_EMPTY) {
+			unsigned int nofs_flags;
+
 			ASSERT(sb_zone_is_full(reset));
 
+			nofs_flags = memalloc_nofs_save();
 			ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
 					       reset->start, reset->len,
-					       GFP_NOFS);
+					       GFP_KERNEL);
+			memalloc_nofs_restore(nofs_flags);
 			if (ret)
 				return ret;
 
@@ -974,11 +978,14 @@ int btrfs_advance_sb_log(struct btrfs_device *device, int mirror)
 			 * explicit ZONE_FINISH is not necessary.
 			 */
 			if (zone->wp != zone->start + zone->capacity) {
+				unsigned int nofs_flags;
 				int ret;
 
+				nofs_flags = memalloc_nofs_save();
 				ret = blkdev_zone_mgmt(device->bdev,
 						REQ_OP_ZONE_FINISH, zone->start,
-						zone->len, GFP_NOFS);
+						zone->len, GFP_KERNEL);
+				memalloc_nofs_restore(nofs_flags);
 				if (ret)
 					return ret;
 			}
@@ -996,11 +1003,13 @@ int btrfs_advance_sb_log(struct btrfs_device *device, int mirror)
 
 int btrfs_reset_sb_log_zones(struct block_device *bdev, int mirror)
 {
+	unsigned int nofs_flags;
 	sector_t zone_sectors;
 	sector_t nr_sectors;
 	u8 zone_sectors_shift;
 	u32 sb_zone;
 	u32 nr_zones;
+	int ret;
 
 	zone_sectors = bdev_zone_sectors(bdev);
 	zone_sectors_shift = ilog2(zone_sectors);
@@ -1011,9 +1020,13 @@ int btrfs_reset_sb_log_zones(struct block_device *bdev, int mirror)
 	if (sb_zone + 1 >= nr_zones)
 		return -ENOENT;
 
-	return blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-				zone_start_sector(sb_zone, bdev),
-				zone_sectors * BTRFS_NR_SB_LOG_ZONES, GFP_NOFS);
+	nofs_flags = memalloc_nofs_save();
+	ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
+			       zone_start_sector(sb_zone, bdev),
+			       zone_sectors * BTRFS_NR_SB_LOG_ZONES,
+			       GFP_KERNEL);
+	memalloc_nofs_restore(nofs_flags);
+	return ret;
 }
 
 /*
@@ -1124,12 +1137,15 @@ static void btrfs_dev_clear_active_zone(struct btrfs_device *device, u64 pos)
 int btrfs_reset_device_zone(struct btrfs_device *device, u64 physical,
 			    u64 length, u64 *bytes)
 {
+	unsigned int nofs_flags;
 	int ret;
 
 	*bytes = 0;
+	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_RESET,
 			       physical >> SECTOR_SHIFT, length >> SECTOR_SHIFT,
-			       GFP_NOFS);
+			       GFP_KERNEL);
+	memalloc_nofs_restore(nofs_flags);
 	if (ret)
 		return ret;
 
@@ -2234,14 +2250,17 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 		struct btrfs_device *device = map->stripes[i].dev;
 		const u64 physical = map->stripes[i].physical;
 		struct btrfs_zoned_device_info *zinfo = device->zone_info;
+		unsigned int nofs_flags;
 
 		if (zinfo->max_active_zones == 0)
 			continue;
 
+		nofs_flags = memalloc_nofs_save();
 		ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_FINISH,
 				       physical >> SECTOR_SHIFT,
 				       zinfo->zone_size >> SECTOR_SHIFT,
-				       GFP_NOFS);
+				       GFP_KERNEL);
+		memalloc_nofs_restore(nofs_flags);
 
 		if (ret)
 			return ret;

-- 
2.43.0

[PATCH v3 4/5] f2fs: guard blkdev_zone_mgmt with nofs scope

[Johannes Thumshirn]

Guard the calls to blkdev_zone_mgmt() with a memalloc_nofs scope.
This helps us getting rid of the GFP_NOFS argument to blkdev_zone_mgmt();

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
---
 fs/f2fs/segment.c | 15 ++++++++++++---
 1 file changed, 12 insertions(+), 3 deletions(-)

diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 4c8836ded90f..0094fe491364 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -1971,9 +1971,15 @@ static int __f2fs_issue_discard_zone(struct f2fs_sb_info *sbi,
 		}
 
 		if (unlikely(is_sbi_flag_set(sbi, SBI_POR_DOING))) {
+			unsigned int nofs_flags;
+			int ret;
+
 			trace_f2fs_issue_reset_zone(bdev, blkstart);
-			return blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-						sector, nr_sects, GFP_NOFS);
+			nofs_flags = memalloc_nofs_save();
+			ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
+						sector, nr_sects, GFP_KERNEL);
+			memalloc_nofs_restore(nofs_flags);
+			return ret;
 		}
 
 		__queue_zone_reset_cmd(sbi, bdev, blkstart, lblkstart, blklen);
@@ -4865,6 +4871,7 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 	block_t zone_block, valid_block_cnt;
 	unsigned int log_sectors_per_block = sbi->log_blocksize - SECTOR_SHIFT;
 	int ret;
+	unsigned int nofs_flags;
 
 	if (zone->type != BLK_ZONE_TYPE_SEQWRITE_REQ)
 		return 0;
@@ -4912,8 +4919,10 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 		    "pointer: valid block[0x%x,0x%x] cond[0x%x]",
 		    zone_segno, valid_block_cnt, zone->cond);
 
+	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(fdev->bdev, REQ_OP_ZONE_FINISH,
-				zone->start, zone->len, GFP_NOFS);
+				zone->start, zone->len, GFP_KERNEL);
+	memalloc_nofs_restore(nofs_flags);
 	if (ret == -EOPNOTSUPP) {
 		ret = blkdev_issue_zeroout(fdev->bdev, zone->wp,
 					zone->len - (zone->wp - zone->start),

-- 
2.43.0

[PATCH v3 5/5] block: remove gfp_flags from blkdev_zone_mgmt

[Johannes Thumshirn]

Now that all callers pass in GFP_KERNEL to blkdev_zone_mgmt() and use
memalloc_no{io,fs}_{save,restore}() to define the allocation scope, we can
drop the gfp_mask parameter from blkdev_zone_mgmt() as well as
blkdev_zone_reset_all() and blkdev_zone_reset_all_emulated().

Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
---
 block/blk-zoned.c              | 19 ++++++++-----------
 drivers/md/dm-zoned-metadata.c |  2 +-
 drivers/nvme/target/zns.c      |  5 ++---
 fs/btrfs/zoned.c               | 14 +++++---------
 fs/f2fs/segment.c              |  4 ++--
 fs/zonefs/super.c              |  2 +-
 include/linux/blkdev.h         |  2 +-
 7 files changed, 20 insertions(+), 28 deletions(-)

diff --git a/block/blk-zoned.c b/block/blk-zoned.c
index d343e5756a9c..d4f4f8325eff 100644
--- a/block/blk-zoned.c
+++ b/block/blk-zoned.c
@@ -177,8 +177,7 @@ static int blk_zone_need_reset_cb(struct blk_zone *zone, unsigned int idx,
 	}
 }
 
-static int blkdev_zone_reset_all_emulated(struct block_device *bdev,
-					  gfp_t gfp_mask)
+static int blkdev_zone_reset_all_emulated(struct block_device *bdev)
 {
 	struct gendisk *disk = bdev->bd_disk;
 	sector_t capacity = bdev_nr_sectors(bdev);
@@ -205,7 +204,7 @@ static int blkdev_zone_reset_all_emulated(struct block_device *bdev,
 		}
 
 		bio = blk_next_bio(bio, bdev, 0, REQ_OP_ZONE_RESET | REQ_SYNC,
-				   gfp_mask);
+				   GFP_KERNEL);
 		bio->bi_iter.bi_sector = sector;
 		sector += zone_sectors;
 
@@ -223,7 +222,7 @@ static int blkdev_zone_reset_all_emulated(struct block_device *bdev,
 	return ret;
 }
 
-static int blkdev_zone_reset_all(struct block_device *bdev, gfp_t gfp_mask)
+static int blkdev_zone_reset_all(struct block_device *bdev)
 {
 	struct bio bio;
 
@@ -238,7 +237,6 @@ static int blkdev_zone_reset_all(struct block_device *bdev, gfp_t gfp_mask)
  * @sector:	Start sector of the first zone to operate on
  * @nr_sectors:	Number of sectors, should be at least the length of one zone and
  *		must be zone size aligned.
- * @gfp_mask:	Memory allocation flags (for bio_alloc)
  *
  * Description:
  *    Perform the specified operation on the range of zones specified by
@@ -248,7 +246,7 @@ static int blkdev_zone_reset_all(struct block_device *bdev, gfp_t gfp_mask)
  *    or finish request.
  */
 int blkdev_zone_mgmt(struct block_device *bdev, enum req_op op,
-		     sector_t sector, sector_t nr_sectors, gfp_t gfp_mask)
+		     sector_t sector, sector_t nr_sectors)
 {
 	struct request_queue *q = bdev_get_queue(bdev);
 	sector_t zone_sectors = bdev_zone_sectors(bdev);
@@ -285,12 +283,12 @@ int blkdev_zone_mgmt(struct block_device *bdev, enum req_op op,
 	 */
 	if (op == REQ_OP_ZONE_RESET && sector == 0 && nr_sectors == capacity) {
 		if (!blk_queue_zone_resetall(q))
-			return blkdev_zone_reset_all_emulated(bdev, gfp_mask);
-		return blkdev_zone_reset_all(bdev, gfp_mask);
+			return blkdev_zone_reset_all_emulated(bdev);
+		return blkdev_zone_reset_all(bdev);
 	}
 
 	while (sector < end_sector) {
-		bio = blk_next_bio(bio, bdev, 0, op | REQ_SYNC, gfp_mask);
+		bio = blk_next_bio(bio, bdev, 0, op | REQ_SYNC, GFP_KERNEL);
 		bio->bi_iter.bi_sector = sector;
 		sector += zone_sectors;
 
@@ -419,8 +417,7 @@ int blkdev_zone_mgmt_ioctl(struct block_device *bdev, blk_mode_t mode,
 		return -ENOTTY;
 	}
 
-	ret = blkdev_zone_mgmt(bdev, op, zrange.sector, zrange.nr_sectors,
-			       GFP_KERNEL);
+	ret = blkdev_zone_mgmt(bdev, op, zrange.sector, zrange.nr_sectors);
 
 fail:
 	if (cmd == BLKRESETZONE)
diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
index 165996cc966c..8156881a31de 100644
--- a/drivers/md/dm-zoned-metadata.c
+++ b/drivers/md/dm-zoned-metadata.c
@@ -1660,7 +1660,7 @@ static int dmz_reset_zone(struct dmz_metadata *zmd, struct dm_zone *zone)
 		noio_flag = memalloc_noio_save();
 		ret = blkdev_zone_mgmt(dev->bdev, REQ_OP_ZONE_RESET,
 				       dmz_start_sect(zmd, zone),
-				       zmd->zone_nr_sectors, GFP_KERNEL);
+				       zmd->zone_nr_sectors);
 		memalloc_noio_restore(noio_flag);
 		if (ret) {
 			dmz_dev_err(dev, "Reset zone %u failed %d",
diff --git a/drivers/nvme/target/zns.c b/drivers/nvme/target/zns.c
index 5b5c1e481722..3148d9f1bde6 100644
--- a/drivers/nvme/target/zns.c
+++ b/drivers/nvme/target/zns.c
@@ -456,8 +456,7 @@ static u16 nvmet_bdev_execute_zmgmt_send_all(struct nvmet_req *req)
 	switch (zsa_req_op(req->cmd->zms.zsa)) {
 	case REQ_OP_ZONE_RESET:
 		ret = blkdev_zone_mgmt(req->ns->bdev, REQ_OP_ZONE_RESET, 0,
-				       get_capacity(req->ns->bdev->bd_disk),
-				       GFP_KERNEL);
+				       get_capacity(req->ns->bdev->bd_disk));
 		if (ret < 0)
 			return blkdev_zone_mgmt_errno_to_nvme_status(ret);
 		break;
@@ -508,7 +507,7 @@ static void nvmet_bdev_zmgmt_send_work(struct work_struct *w)
 		goto out;
 	}
 
-	ret = blkdev_zone_mgmt(bdev, op, sect, zone_sectors, GFP_KERNEL);
+	ret = blkdev_zone_mgmt(bdev, op, sect, zone_sectors);
 	if (ret < 0)
 		status = blkdev_zone_mgmt_errno_to_nvme_status(ret);
 
diff --git a/fs/btrfs/zoned.c b/fs/btrfs/zoned.c
index 05640d61e435..cf2e779d8ef4 100644
--- a/fs/btrfs/zoned.c
+++ b/fs/btrfs/zoned.c
@@ -830,8 +830,7 @@ static int sb_log_location(struct block_device *bdev, struct blk_zone *zones,
 
 			nofs_flags = memalloc_nofs_save();
 			ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-					       reset->start, reset->len,
-					       GFP_KERNEL);
+					       reset->start, reset->len);
 			memalloc_nofs_restore(nofs_flags);
 			if (ret)
 				return ret;
@@ -984,7 +983,7 @@ int btrfs_advance_sb_log(struct btrfs_device *device, int mirror)
 				nofs_flags = memalloc_nofs_save();
 				ret = blkdev_zone_mgmt(device->bdev,
 						REQ_OP_ZONE_FINISH, zone->start,
-						zone->len, GFP_KERNEL);
+						zone->len);
 				memalloc_nofs_restore(nofs_flags);
 				if (ret)
 					return ret;
@@ -1023,8 +1022,7 @@ int btrfs_reset_sb_log_zones(struct block_device *bdev, int mirror)
 	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
 			       zone_start_sector(sb_zone, bdev),
-			       zone_sectors * BTRFS_NR_SB_LOG_ZONES,
-			       GFP_KERNEL);
+			       zone_sectors * BTRFS_NR_SB_LOG_ZONES);
 	memalloc_nofs_restore(nofs_flags);
 	return ret;
 }
@@ -1143,8 +1141,7 @@ int btrfs_reset_device_zone(struct btrfs_device *device, u64 physical,
 	*bytes = 0;
 	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_RESET,
-			       physical >> SECTOR_SHIFT, length >> SECTOR_SHIFT,
-			       GFP_KERNEL);
+			       physical >> SECTOR_SHIFT, length >> SECTOR_SHIFT);
 	memalloc_nofs_restore(nofs_flags);
 	if (ret)
 		return ret;
@@ -2258,8 +2255,7 @@ static int do_zone_finish(struct btrfs_block_group *block_group, bool fully_writ
 		nofs_flags = memalloc_nofs_save();
 		ret = blkdev_zone_mgmt(device->bdev, REQ_OP_ZONE_FINISH,
 				       physical >> SECTOR_SHIFT,
-				       zinfo->zone_size >> SECTOR_SHIFT,
-				       GFP_KERNEL);
+				       zinfo->zone_size >> SECTOR_SHIFT);
 		memalloc_nofs_restore(nofs_flags);
 
 		if (ret)
diff --git a/fs/f2fs/segment.c b/fs/f2fs/segment.c
index 0094fe491364..e1065ba70207 100644
--- a/fs/f2fs/segment.c
+++ b/fs/f2fs/segment.c
@@ -1977,7 +1977,7 @@ static int __f2fs_issue_discard_zone(struct f2fs_sb_info *sbi,
 			trace_f2fs_issue_reset_zone(bdev, blkstart);
 			nofs_flags = memalloc_nofs_save();
 			ret = blkdev_zone_mgmt(bdev, REQ_OP_ZONE_RESET,
-						sector, nr_sects, GFP_KERNEL);
+						sector, nr_sects);
 			memalloc_nofs_restore(nofs_flags);
 			return ret;
 		}
@@ -4921,7 +4921,7 @@ static int check_zone_write_pointer(struct f2fs_sb_info *sbi,
 
 	nofs_flags = memalloc_nofs_save();
 	ret = blkdev_zone_mgmt(fdev->bdev, REQ_OP_ZONE_FINISH,
-				zone->start, zone->len, GFP_KERNEL);
+				zone->start, zone->len);
 	memalloc_nofs_restore(nofs_flags);
 	if (ret == -EOPNOTSUPP) {
 		ret = blkdev_issue_zeroout(fdev->bdev, zone->wp,
diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
index 63fbac018c04..cadb1364f951 100644
--- a/fs/zonefs/super.c
+++ b/fs/zonefs/super.c
@@ -113,7 +113,7 @@ static int zonefs_zone_mgmt(struct super_block *sb,
 
 	trace_zonefs_zone_mgmt(sb, z, op);
 	ret = blkdev_zone_mgmt(sb->s_bdev, op, z->z_sector,
-			       z->z_size >> SECTOR_SHIFT, GFP_KERNEL);
+			       z->z_size >> SECTOR_SHIFT);
 	if (ret) {
 		zonefs_err(sb,
 			   "Zone management operation %s at %llu failed %d\n",
diff --git a/include/linux/blkdev.h b/include/linux/blkdev.h
index 99e4f5e72213..8467c1910404 100644
--- a/include/linux/blkdev.h
+++ b/include/linux/blkdev.h
@@ -325,7 +325,7 @@ void disk_set_zoned(struct gendisk *disk);
 int blkdev_report_zones(struct block_device *bdev, sector_t sector,
 		unsigned int nr_zones, report_zones_cb cb, void *data);
 int blkdev_zone_mgmt(struct block_device *bdev, enum req_op op,
-		sector_t sectors, sector_t nr_sectors, gfp_t gfp_mask);
+		sector_t sectors, sector_t nr_sectors);
 int blk_revalidate_disk_zones(struct gendisk *disk,
 		void (*update_driver_data)(struct gendisk *disk));
 

-- 
2.43.0

Re: [PATCH v3 1/5] zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call

[Damien Le Moal]

On 1/29/24 16:52, Johannes Thumshirn wrote:
> Pass GFP_KERNEL instead of GFP_NOFS to the blkdev_zone_mgmt() call in
> zonefs_zone_mgmt().
> 
> As as zonefs_zone_mgmt() and zonefs_inode_zone_mgmt() are never called
> from a place that can recurse back into the filesystem on memory reclaim,
> it is save to call blkdev_zone_mgmt() with GFP_KERNEL.

s/save/safe

> Link: https://lore.kernel.org/all/ZZcgXI46AinlcBDP@casper.infradead.org/
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
> ---
>  fs/zonefs/super.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
> index 93971742613a..63fbac018c04 100644
> --- a/fs/zonefs/super.c
> +++ b/fs/zonefs/super.c
> @@ -113,7 +113,7 @@ static int zonefs_zone_mgmt(struct super_block *sb,
>  
>  	trace_zonefs_zone_mgmt(sb, z, op);
>  	ret = blkdev_zone_mgmt(sb->s_bdev, op, z->z_sector,
> -			       z->z_size >> SECTOR_SHIFT, GFP_NOFS);
> +			       z->z_size >> SECTOR_SHIFT, GFP_KERNEL);
>  	if (ret) {
>  		zonefs_err(sb,
>  			   "Zone management operation %s at %llu failed %d\n",

I think this is OK but I will need a little more time to fully convince myself.
So let me look again at the code to check all the calls contexts.

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH v3 2/5] dm: dm-zoned: guard blkdev_zone_mgmt with noio scope

[Damien Le Moal]

On 1/29/24 16:52, Johannes Thumshirn wrote:
> Guard the calls to blkdev_zone_mgmt() with a memalloc_noio scope.
> This helps us getting rid of the GFP_NOIO argument to blkdev_zone_mgmt();
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

> ---
>  drivers/md/dm-zoned-metadata.c | 5 ++++-
>  1 file changed, 4 insertions(+), 1 deletion(-)
> 
> diff --git a/drivers/md/dm-zoned-metadata.c b/drivers/md/dm-zoned-metadata.c
> index fdfe30f7b697..165996cc966c 100644
> --- a/drivers/md/dm-zoned-metadata.c
> +++ b/drivers/md/dm-zoned-metadata.c
> @@ -1655,10 +1655,13 @@ static int dmz_reset_zone(struct dmz_metadata *zmd, struct dm_zone *zone)
>  
>  	if (!dmz_is_empty(zone) || dmz_seq_write_err(zone)) {
>  		struct dmz_dev *dev = zone->dev;
> +		unsigned int noio_flag;
>  
> +		noio_flag = memalloc_noio_save();
>  		ret = blkdev_zone_mgmt(dev->bdev, REQ_OP_ZONE_RESET,
>  				       dmz_start_sect(zmd, zone),
> -				       zmd->zone_nr_sectors, GFP_NOIO);
> +				       zmd->zone_nr_sectors, GFP_KERNEL);
> +		memalloc_noio_restore(noio_flag);
>  		if (ret) {
>  			dmz_dev_err(dev, "Reset zone %u failed %d",
>  				    zone->id, ret);
> 

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH v3 3/5] btrfs: zoned: call blkdev_zone_mgmt in nofs scope

[Damien Le Moal]

On 1/29/24 16:52, Johannes Thumshirn wrote:
> Add a memalloc_nofs scope around all calls to blkdev_zone_mgmt(). This
> allows us to further get rid of the GFP_NOFS argument for
> blkdev_zone_mgmt().
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Looks OK to me.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>


-- 
Damien Le Moal
Western Digital Research

Re: [PATCH v3 4/5] f2fs: guard blkdev_zone_mgmt with nofs scope

[Damien Le Moal]

On 1/29/24 16:52, Johannes Thumshirn wrote:
> Guard the calls to blkdev_zone_mgmt() with a memalloc_nofs scope.
> This helps us getting rid of the GFP_NOFS argument to blkdev_zone_mgmt();
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH v3 5/5] block: remove gfp_flags from blkdev_zone_mgmt

[Damien Le Moal]

On 1/29/24 16:52, Johannes Thumshirn wrote:
> Now that all callers pass in GFP_KERNEL to blkdev_zone_mgmt() and use
> memalloc_no{io,fs}_{save,restore}() to define the allocation scope, we can
> drop the gfp_mask parameter from blkdev_zone_mgmt() as well as
> blkdev_zone_reset_all() and blkdev_zone_reset_all_emulated().
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Looks good to me.

Reviewed-by: Damien Le Moal <dlemoal@kernel.org>

But let me check zonefs (patch 1).

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH v3 2/5] dm: dm-zoned: guard blkdev_zone_mgmt with noio scope

[Mike Snitzer]

On Mon, Jan 29 2024 at  2:52P -0500,
Johannes Thumshirn <johannes.thumshirn@wdc.com> wrote:

> Guard the calls to blkdev_zone_mgmt() with a memalloc_noio scope.
> This helps us getting rid of the GFP_NOIO argument to blkdev_zone_mgmt();
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Reviewed-by: Mike Snitzer <snitzer@kernel.org>

Re: [PATCH v3 5/5] block: remove gfp_flags from blkdev_zone_mgmt

[Mike Snitzer]

On Mon, Jan 29 2024 at  2:52P -0500,
Johannes Thumshirn <johannes.thumshirn@wdc.com> wrote:

> Now that all callers pass in GFP_KERNEL to blkdev_zone_mgmt() and use
> memalloc_no{io,fs}_{save,restore}() to define the allocation scope, we can
> drop the gfp_mask parameter from blkdev_zone_mgmt() as well as
> blkdev_zone_reset_all() and blkdev_zone_reset_all_emulated().
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Reviewed-by: Mike Snitzer <snitzer@kernel.org>

Re: [PATCH v3 3/5] btrfs: zoned: call blkdev_zone_mgmt in nofs scope

[David Sterba]

On Sun, Jan 28, 2024 at 11:52:18PM -0800, Johannes Thumshirn wrote:
> Add a memalloc_nofs scope around all calls to blkdev_zone_mgmt(). This
> allows us to further get rid of the GFP_NOFS argument for
> blkdev_zone_mgmt().
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Reviewed-by: David Sterba <dsterba@suse.com>

Re: [PATCH v3 4/5] f2fs: guard blkdev_zone_mgmt with nofs scope

[Chao Yu]

On 2024/1/29 15:52, Johannes Thumshirn wrote:
> Guard the calls to blkdev_zone_mgmt() with a memalloc_nofs scope.
> This helps us getting rid of the GFP_NOFS argument to blkdev_zone_mgmt();
> 
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>

Reviewed-by: Chao Yu <chao@kernel.org>

Thanks,

Re: [PATCH v3 1/5] zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call

[Damien Le Moal]

On 1/29/24 16:52, Johannes Thumshirn wrote:
> Pass GFP_KERNEL instead of GFP_NOFS to the blkdev_zone_mgmt() call in
> zonefs_zone_mgmt().
> 
> As as zonefs_zone_mgmt() and zonefs_inode_zone_mgmt() are never called
> from a place that can recurse back into the filesystem on memory reclaim,
> it is save to call blkdev_zone_mgmt() with GFP_KERNEL.
> 
> Link: https://lore.kernel.org/all/ZZcgXI46AinlcBDP@casper.infradead.org/
> Signed-off-by: Johannes Thumshirn <johannes.thumshirn@wdc.com>
> ---
>  fs/zonefs/super.c | 2 +-
>  1 file changed, 1 insertion(+), 1 deletion(-)
> 
> diff --git a/fs/zonefs/super.c b/fs/zonefs/super.c
> index 93971742613a..63fbac018c04 100644
> --- a/fs/zonefs/super.c
> +++ b/fs/zonefs/super.c
> @@ -113,7 +113,7 @@ static int zonefs_zone_mgmt(struct super_block *sb,
>  
>  	trace_zonefs_zone_mgmt(sb, z, op);
>  	ret = blkdev_zone_mgmt(sb->s_bdev, op, z->z_sector,
> -			       z->z_size >> SECTOR_SHIFT, GFP_NOFS);
> +			       z->z_size >> SECTOR_SHIFT, GFP_KERNEL);
>  	if (ret) {
>  		zonefs_err(sb,
>  			   "Zone management operation %s at %llu failed %d\n",
> 

Given that zonefs_inode_zone_mgmt() which calls zonefs_zone_mgmt() is only used
for sequential zone inodes, and that these inodes cannot be written with
buffered IOs, I think this is safe as we will never have dirty pages to
writeback for reclaim. So there should be no risk of re-entering the FS on
reclaim with GFP_KERNEL.

So:

Acked-by: Damien Le Moal <dlemoal@kernel.org>

-- 
Damien Le Moal
Western Digital Research

Re: [PATCH v3 0/5] block: remove gfp_mask for blkdev_zone_mgmt()

[Johannes Thumshirn]

On 29.01.24 08:52, Johannes Thumshirn wrote:
> Fueled by the LSFMM discussion on removing GFP_NOFS initiated by Willy,
> I've looked into the sole GFP_NOFS allocation in zonefs. As it turned out,
> it is only done for zone management commands and can be removed.
> 
> After digging into more callers of blkdev_zone_mgmt() I came to the
> conclusion that the gfp_mask parameter can be removed alltogether.
> 
> So this series switches all callers of blkdev_zone_mgmt() to either use
> GFP_KERNEL where possible or grab a memalloc_no{fs,io} context.
> 
> The final patch in this series is getting rid of the gfp_mask parameter.
> 
> Link: https://lore.kernel.org/all/ZZcgXI46AinlcBDP@casper.infradead.org/
> 
> ---
> Changes in v3:
> - Fix build error after rebase in dm-zoned-metadata.c
> - Link to v2: https://lore.kernel.org/r/20240125-zonefs_nofs-v2-0-2d975c8c1690@wdc.com
> 
> Changes in v2:
> - guard blkdev_zone_mgmt in dm-zoned-metadata.c with memalloc_noio context
> - Link to v1: https://lore.kernel.org/r/20240123-zonefs_nofs-v1-0-cc0b0308ef25@wdc.com
> 
> ---
> Johannes Thumshirn (5):
>        zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call
>        dm: dm-zoned: guard blkdev_zone_mgmt with noio scope
>        btrfs: zoned: call blkdev_zone_mgmt in nofs scope
>        f2fs: guard blkdev_zone_mgmt with nofs scope
>        block: remove gfp_flags from blkdev_zone_mgmt
> 
>   block/blk-zoned.c              | 19 ++++++++-----------
>   drivers/md/dm-zoned-metadata.c |  5 ++++-
>   drivers/nvme/target/zns.c      |  5 ++---
>   fs/btrfs/zoned.c               | 35 +++++++++++++++++++++++++----------
>   fs/f2fs/segment.c              | 15 ++++++++++++---
>   fs/zonefs/super.c              |  2 +-
>   include/linux/blkdev.h         |  2 +-
>   7 files changed, 53 insertions(+), 30 deletions(-)
> ---
> base-commit: 615d300648869c774bd1fe54b4627bb0c20faed4
> change-id: 20240110-zonefs_nofs-dd1e22b2e046
> 
> Best regards,

Hi Jens,

now that every patch is reviewed, can you queue this series please?

Thanks,
Johannes

Re: [PATCH v3 0/5] block: remove gfp_mask for blkdev_zone_mgmt()

[Jens Axboe]

On Sun, 28 Jan 2024 23:52:15 -0800, Johannes Thumshirn wrote:
> Fueled by the LSFMM discussion on removing GFP_NOFS initiated by Willy,
> I've looked into the sole GFP_NOFS allocation in zonefs. As it turned out,
> it is only done for zone management commands and can be removed.
> 
> After digging into more callers of blkdev_zone_mgmt() I came to the
> conclusion that the gfp_mask parameter can be removed alltogether.
> 
> [...]

Applied, thanks!

[1/5] zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call
      commit: 9105ce591b424771b1502ef9836ca7953c3e0af4
[2/5] dm: dm-zoned: guard blkdev_zone_mgmt with noio scope
      commit: 218082010aceb40b5495ebc30028ede6e30ee755
[3/5] btrfs: zoned: call blkdev_zone_mgmt in nofs scope
      commit: d9d556755f16f6af8d1d8ebac38b83a9263394c5
[4/5] f2fs: guard blkdev_zone_mgmt with nofs scope
      commit: 147ec1c60e3273d21ea1f212c6636f231d6d2771
[5/5] block: remove gfp_flags from blkdev_zone_mgmt
      commit: 71f4ecdbb42addf82b01b734b122a02707fed521

Best regards,
-- 
Jens Axboe

Re: [f2fs-dev] [PATCH v3 0/5] block: remove gfp_mask for blkdev_zone_mgmt()

[patchwork-bot+f2fs]

Hello:

This series was applied to jaegeuk/f2fs.git (dev)
by Jens Axboe <axboe@kernel.dk>:

On Sun, 28 Jan 2024 23:52:15 -0800 you wrote:
> Fueled by the LSFMM discussion on removing GFP_NOFS initiated by Willy,
> I've looked into the sole GFP_NOFS allocation in zonefs. As it turned out,
> it is only done for zone management commands and can be removed.
> 
> After digging into more callers of blkdev_zone_mgmt() I came to the
> conclusion that the gfp_mask parameter can be removed alltogether.
> 
> [...]

Here is the summary with links:
  - [f2fs-dev,v3,1/5] zonefs: pass GFP_KERNEL to blkdev_zone_mgmt() call
    https://git.kernel.org/jaegeuk/f2fs/c/9105ce591b42
  - [f2fs-dev,v3,2/5] dm: dm-zoned: guard blkdev_zone_mgmt with noio scope
    https://git.kernel.org/jaegeuk/f2fs/c/218082010ace
  - [f2fs-dev,v3,3/5] btrfs: zoned: call blkdev_zone_mgmt in nofs scope
    https://git.kernel.org/jaegeuk/f2fs/c/d9d556755f16
  - [f2fs-dev,v3,4/5] f2fs: guard blkdev_zone_mgmt with nofs scope
    https://git.kernel.org/jaegeuk/f2fs/c/147ec1c60e32
  - [f2fs-dev,v3,5/5] block: remove gfp_flags from blkdev_zone_mgmt
    https://git.kernel.org/jaegeuk/f2fs/c/71f4ecdbb42a

You are awesome, thank you!
-- 
Deet-doot-dot, I am a bot.
https://korg.docs.kernel.org/patchwork/pwbot.html